CPSC 322, Lecture 17Slide 1 Planning: Representation and Forward Search Computer Science cpsc322,...

CPSC 322, Lecture 17 Slide 1

Planning: Representation and

Forward Search Computer Science cpsc322, Lecture 17

(Textbook Chpt 8.1 (Skip 8.1.1-2)- 8.2)

February, 10, 2010

Course Announcements• Final Exam Apr 19, 12:00 pm (3 hours, DMP

201  )• Solutions for ArcC practice ex. + practice ex.

for SLS posted• Assign2 on CSPs will be posted after class

(due March 3)

CPSC 322, Lecture 17 Slide 2

Subjects needed for NLP experiment.• Tomorrow and Fri, flexible schedule. 2-2.5h ->

20$• If interested send email asap to:

gabriel.murray@gmail.com

CPSC 322, Lecture 17 Slide 3

Lecture Overview

• Clarification• Where are we?• Planning

• Example• STRIPS: a Feature-Based Representation• Forward Planning

CPSC 322, Lecture 16 Slide 4

Sampling a discrete probability distribution

CPSC 322, Lecture 17 Slide 5

Lecture Overview

• Clarifications• Where are we?• Planning

• Example• STRIPS: a Feature-Based Representation• Forward Planning

CPSC 322, Lecture 2 Slide 6

Modules we'll cover in this course: R&Rsys

Environment

Problem

Inference

Planning

Deterministic

Stochastic

SearchArc Consistency

Search

Search Value Iteration

Var. Elimination

Constraint Satisfactio

n

Logics

STRIPS

Belief Nets

Vars + Constraint

s

Decision Nets

Markov Processes

Var. Elimination

Static

Sequential

RepresentationReasoningTechnique

SLS

CPSC 322, Lecture 11 Slide 7

Standard Search vs. Specific R&R systemsConstraint Satisfaction (Problems):

• State: assignments of values to a subset of the variables• Successor function: assign values to a “free” variable• Goal test: set of constraints• Solution: possible world that satisfies the constraints• Heuristic function: none (all solutions at the same distance from

start)

Planning : • State• Successor function• Goal test• Solution• Heuristic function

Inference• State• Successor function• Goal test• Solution• Heuristic function

CPSC 322, Lecture 17 Slide 8

Lecture Overview

• Clarifications• Where are we?• Planning

• Example• STRIPS representation and assumption• Forward Planning

CPSC 322, Lecture 17 Slide 9

Planning as Search: State and Goal

How to select and organize a sequence of actions to achieve a given goal…

State: Agent is in a possible world (full assignments to a set of variables/features)

Goal: Agent wants to be in a possible world were some variables are given specific values

CPSC 322, Lecture 17 Slide 10

Planning as Search: Successor function and Solution

Actions : take the agent from one state to another

Solution: sequence of actions that when performed will take the agent from the current state to a goal state

CPSC 322, Lecture 17 Slide 11

Lecture Overview

• Clarifications• Where are we?• Planning

• Example• STRIPS representation and assumption• Forward Planning

CPSC 322, Lecture 17 Slide 12

Delivery Robot Example (textbook)

Consider a delivery robot named Rob, who must navigate the following environment, can deliver coffee and mail to Sam

Another example will be available as a Practice Exercise:

“Commuting to UBC”

CPSC 322, Lecture 17 Slide 13

Delivery Robot Example: States

The state is defined by the following variables/features:

RLoc - Rob's location• domain: coffee shop (cs), Sam's office (off ), mail room

(mr ),

or laboratory (lab)

RHC - Rob has coffee True/False. SWC - Sam wants coffeeMW - Mail is waitingRHM - Rob has mail

rhcrhc

Example state:Number of states:

CPSC 322, Lecture 17 Slide 14

Delivery Robot Example:Actions

The robot’s actions are:Move - Rob's move action

• move clockwise (mc ), move anti-clockwise (mac ) not move (nm )

PUC - Rob picks up coffee• must be at the coffee shop

DelC - Rob delivers coffee• must be at the office, and must have coffee

PUM - Rob picks up mail• must be in the mail room, and mail must be

waitingDelM - Rob delivers mail

• must be at the office and have mail

CPSC 322, Lecture 17 Slide 15

Lecture Overview

• Clarifications• Were are we?• Planning

• Example• STRIPS representation and assumption

(STanford Research Institute Problem Solver )

• Forward Planning

CPSC 322, Lecture 17 Slide 16

STRIPS action representation

The key to sophisticated planning is modeling actions

In STRIPS, an action has two parts:1. Preconditions: a set of assignments to features that

must be satisfied in order for the action to be legal2. Effects: a set of assignments to features that are

caused by the action

CPSC 322, Lecture 17 Slide 17

STRIPS actions: Example

STRIPS representation of the action pick up coffee, PUC :• preconditions Loc = cs and RHC = F • effects RHC = T

STRIPS representation of the action deliver coffee, DelC :• preconditions Loc = and RHC = • effects RHC = and SWC =

Note in this domain Sam doesn't have to want coffee for Rob to deliver it; one way or another, Sam doesn't want coffee after delivery.

CPSC 322, Lecture 17 Slide 18

STRIPS actions: MC and MAC

STRIPS representation of the action MoveClockwise ?

CPSC 322, Lecture 17 Slide 19

STRIPS Actions (cont’)

• The STRIPS assumption:• all variables not explicitly changed by an action

stay unchanged

• So if the feature/variable V has value v after the action a has been performed, what can we conclude about a and/or the state of the world immediately preceding the execution of a?

CPSC 322, Lecture 17 Slide 20

Lecture Overview

• Clarifications• Where are we?• Planning

• Example• STRIPS representation and assumption

(STanford Research Institute Problem Solver )

• Forward Planning

CPSC 322, Lecture 17 Slide 21

Forward Planning

To find a plan, a solution: search in the state-space graph.• The states are the possible worlds• The arcs correspond to the actions: The

arcs from a state s represent all of the actions that are legal in state s. (What actions are legal?)

• A plan is a path from the state representing the initial state to a state that satisfies the goal.

CPSC 322, Lecture 17 Slide 22

Example state-space graph: first level

CPSC 322, Lecture 17 Slide 23

Example state-space

graph

CPSC 322, Lecture 4 Slide 24

Learning Goals for today’s class

You can:

• Represent a planning problem with the STRIPS representation

• Explain the STRIPS assumption • Solve a planning problem by search

(forward planning). Specify states, successor function, goal test and solution.

CPSC 322, Lecture 17 Slide 25

Next class

Finish Planning (Chp 8)• Heuristics for planning (not on textbook)• Mapping planning problem into a CSP (8.4)

CPSC 322, Lecture 13 Slide 26

Feedback summary • Assignments (programming, unclear) 10 1 12

(-2)

• Practice Exercises (too easy) 6 1 4 (+2)

• TAs 6 0 0 (+6)

• Lectures (more interactive) 12 7 6 (+8)

• Course Topics (Bayesian Nets?) 9 1 0 (+8)

• Learning Goals 10 0 0 (+10)

• Textbook 14 0 6(+12)

• Slides(hard to read) 17 2 2 (+15)

• AIspace 17 0 2 (+15)

CPSC 322, Lecture 16 Slide 27

Feedback specific suggestions (>2 people)

• Post precise due textbook readings• Sync slides right before lecture• Provide reading list of recent research

papers• Use clickers